The Institute of Flight Mechanics and Controls at the University of Stuttgart conducts research into the challenges faced when programming aircraft to fly autonomously
Under current arrangements, pilots have full responsibility for aircraft while flying, operating all systems by taking in information, planning actions and carrying them out. Actually steering the airplane is usually carried out by electrical and mechanical systems operated by the pilot. The next step in this process is fairly obvious: It would be advantageous if an aircraft could monitor and steer itself by designing systems that can communicate with one another directly. The aircraft would then refuse to go beyond its safe operating range, and in critical situations it would withdraw control back from pilots to protect itself. For many years, the University of Stuttgart has been researching in the field of automation of aeronautic devices used in civil aviation. Vincenz Frenzel, scientific assistant at the Institute of Flight Mechanics and Control at the University of Stuttgart explains which ideas remain a vision, and which have already become reality. Frenzel was also a guest speaker at the first #techourfuture event.
According to one definition, autonomy is a “state of self-determination […] and the freedom to make decisions.” In the case of aeronautics, it is especially important that this is not the case. Nothing unforeseen should happen, system must not act independently, and artificial intelligence does not have to get involved. A more fitting definition for such a device would be “highly automated,” since the actions of aeronautic devices should be based on predefined rules and algorithms, that should be entirely reliable and carried out by certified systems.
With airplanes, such systems are described as “refuse to crash” – or “systems in command” rather than the “pilot in command.” The idea of such systems is to compensate for certain human weaknesses: distractions, fatigue, illness, emotions, insufficient accuracy or optimization, or poor decision-making. In essence, many of-the-shelf drones already fly by themselves and prevent themselves from crashing. Drone pilots merely issue commands or set waypoints, and drones fly from point to point without getting themselves into danger.
WHY DEVICES ARE AUTOMATED
There are a number of different motivations for automation depending on the aircraft type. People want to automate air taxis to cut the cost of travel. By replacing the pilot with an additional passenger, payment of crew salaries can be avoided. Depending on the aircraft configuration, almost twice the payload can be transported (freight or passengers) at a lower cost. Often, it is actually not even possible to steer an air taxi manually if it has a large number of engines or control panels, so that a control system is required anyway.
In General Aviation, i.e. private and commercial flying not including airlines and charter flights, the aim of automation is to reduce the requirements placed on pilots and for pilot training. This can improve aircraft sales and enhance the safety of private aviation.
Airline traffic today has a high degree of automation since pilots already have a large number of support systems at their disposal. But because the main cause of airline accidents is still human error, there are developments to further reduce the number of pilots on board aircraft – for reasons of safety, pilot shortages, and economy. The first area in which this will be tried out is air cargo, to avoid posing threats to airline passengers.
Automated flying moved beyond mere technological issues some time ago. This was made possible by the increasing use of fly-by-wire systems, where signals are transmitted in cables wires to control aircraft electronically. This technology has been in use in the airline industry for decades. To achieve even higher levels of automation and extend to new areas of application, the technology will need expanding, however, and certification will be required. By that point, legislators can also address the legal questions that are still open.
THE REQUIREMENTS FOR AUTOMATED AIRCRAFT SYSTEMS
Before introducing newly developed automation systems, it has to be ensured that they are reliable. If necessary, safety requirements have to be revised and then certified. In addition, the right infrastructure has to be put in place for communicating with air-traffic controllers. In uncontrolled airspace, new monitoring systems – such as Detect And Avoid systems – will identify obstacles and prevent aircraft from endangering one another. In an emergency, it must be possible to steer these airplanes from the ground, which in itself poses new challenges.
WORK ON AIRCRAFT AUTOMATION AT THE UNIVERSITY OF STUTTGART
In 2015, a fully automated aircraft was showcased at the University of Stuttgart as part of a project called Fly Smart, itself part of LuFo IV, the federal aeronautics research program. The mission of the aircraft, which was classified under EASA class CS-23, was also to complete a fully autonomous take-off and landing without the support of conventional ground-based navigation.
Specialists at the University of Stuttgart are also coordinating test sites for energy-efficient, electric, and autonomous aircraft in Baden-Wuerttemberg. The first test flights took place in 2019 and further comprehensive research is planned for 2020. The project is being spearheaded by the Institute of Flight Mechanics and Control. The Institutes of Aircraft Design, Aircraft Systems, and Navigation at the Faculty of Aerospace Engineering and Geodesy at the University of Stuttgart are also involved in the project, as are a number of other partners from industry, such as Volocopter and Thales. The test site is receiving €1.3 million of funding from the Baden-Wuerttemberg Ministry of Economic Affairs, Labor, and Housing and its aim is to prove and demonstrate new technologies and concepts.
The researchers are not without their visions: “Within decades, you might be sitting happily in the front row, where the cockpit used to be, soaking in the view. Until that point, you’ll simply rely on updates from the flight deck,” predicts Vincenz Frenzel.
Contact
Vincenz Frenzel (author)
Research assistant
Institute of Flight Mechanics and Controls University of Stuttgart (Stuttgart)